Home
Column
Overview
Embark on a melodic journey through the evolution of electronic dance
music as we delve into the dynamic soundscapes of two iconic artists:
Avicii and Kygo. In this captivating exploration, we dissect the
metamorphosis of musical style over the years, meticulously analyzing
the sonic innovations and creative ingenuity that have defined the
discographies of these electronic titans.
Why Avicii and Kygo,
you may wonder? Beyond their chart-topping hits and global acclaim,
these virtuosos hold a special place in the heart of the researcher.
Year after year, their captivating compositions have dominated the
coveted slots of the Spotify Wrapped Top 5 Artists, a testament not only
to their immense popularity but also to the profound connection forged
between their music and the listener’s soul.
Prepare to be
entranced as we unravel the intricacies of Avicii’s groundbreaking
melodies, tracing the evolution from his euphoric anthems to the
poignant depths of his later works. Contrast this with the
tropical-infused beats and sun-kissed vibes that hallmark Kygo’s
signature sound, a testament to his ability to transport listeners to
exotic locales through the power of music.
Through meticulous
analysis and keen observation, we aim to unravel the mysteries behind
the ever-changing tapestry of sound, exploring how Avicii and Kygo have
sculpted and redefined the landscape of electronic music. Join us on
this exhilarating expedition, where passion meets precision, and every
beat resonates with the pulse of innovation.The start can be found on
the right, with a quick background of the two artists.
Column
The story of Avicii
Tim Bergling (8 September 1989 – 20 April 2018), known professionally
as Avicii, was a Swedish musician, DJ, remixer, and record producer.
At 16, Bergling began posting his remixes on electronic music
forums, which led to his first record deal. He rose to prominence in
2011 with his single “Levels”. His debut studio album, True (2013),
blended electronic music with elements of multiple genres and received
generally positive reviews. It peaked in the top ten in more than
fifteen countries and topped international dance charts; the lead
single, “Wake Me Up”, topped most music markets in Europe and reached
number four in the United States.
In 2015, Bergling released
his second studio album, Stories, and in 2017 he released an EP, Avīci
(01).His catalogue also included the singles “I Could Be the One” with
Nicky Romero, “You Make Me”, “X You”, “Hey Brother”, “Addicted to You”,
“The Days”, “The Nights”, “Waiting for Love”, “Without You” and “Lonely
Together”. Bergling was nominated for a Grammy Award for his work on
“Sunshine” with David Guetta in 2012 and “Levels” in 2013. Several music
publications credit Bergling as among the DJs who ushered electronic
music into Top 40 radio in the early 2010s.
Bergling retired
from touring in 2016 due to health problems, having suffered stress and
poor mental health for several years. On 20 April 2018, Avicii was found
dead in Muscat, Oman. On 1 May, it was reported that the cause of his
death was suicide from self-inflicted injuries from a broken wine
bottle. He was buried on 8 June in his hometown of Stockholm,
Sweden.
(https://simple.wikipedia.org/wiki/Avicii)
The story of Kygo
Kyrre Gørvell-Dahll (born 11 September 1991), known professionally as
Kygo, is a Norwegian DJ and record producer. He garnered international
attention with his December 2013 remix of the track “I See Fire” by Ed
Sheeran and his December 2014 single “Firestone” (featuring Conrad
Sewell). In late 2015, he reached 1 billion streams on Spotify faster
than any other artist, and by June 2016, he had reached 2 billion
streams.
Kygo has since released several singles, such as
“Stole the Show”, “Here for You”, and “Stay” in 2015, “It Ain’t Me”
collaborating with Selena Gomez in 2017 and a 2019 remix of “Higher
Love” performed by Whitney Houston, all of which have debuted on several
international charts. His debut album, Cloud Nine, was released on 13
May 2016.
Kygo became the first house music producer to perform
at an Olympics closing ceremony in August 2016 at the 2016 Rio Olympics.
In March 2018, Billboard ranked Kygo 3rd on its 2018 ranking of dance
musicians titled Billboard Dance 100. The same year, he was ranked 32 on
DJ Mag’s top 100 DJs of the world, and in October 2019 he was ranked 42.
(https://en.wikipedia.org/wiki/Kygo)
The Corpus
Column
About the corpus
The corpus exists of two playlists; one of each artist. I picked all
the songs that are in their albums and most of the singles to create
these playlists. I did not include remixes, or songs where the artist
took part of, but is not featured in it. In the table on the right are
the most interesting statistics listed and compared. The playlist with
Avicii’s songs can be found here.
The playlist with Kygo’s songs can be found here.
On the right side, a few box plots can be seen with a more
detailed comparison of the two playlists, where five different aspects
are treated. A definition of the features that will be discussed in this
report can be found on the Spotify Documentation (or this
link).
In the first box plot, the energy of the songs can be
seen. Here, you can see that the songs of Avicii have a somewhat higher
energy. The median of both plots is almost right at the middle, so the
plots are both normally distributed. One thing that can be noticed is
the one outlier in Kygo’s plot. Considering the results in the upcoming
tabs, it can be concluded that this outlier is the song ‘Intro’.
In the danceability plot can be seen that the songs of Avicii have
a much wider range of songs in terms of danceability. The plots are,
again, normally distributed. This time, Avicii’s playlist has one
outlier, while Kygo has three of them.
If we look at the next
plot, the acousticness, we can conclude some interesting things. The
most obvious one is that Avicii’s songs have a really low acousticness.
This should not be a surprise really, as he is know as a famous EDM
(Electronic Dance Music) artist. Kygo has a higher acousticness, but
still not a lot of songs above 0.4. Even the outliers of Kygo are a lot
higher than Avicii’s. Because of the median, which is close to the
bottom of the box in Avicii’s plot, this plot is positive skewed. This
highlights the fact that most of the songs have a low acousticness.
The next plot shows the tempo of the songs. Interesting to notice
is the negative skew of Avicii’s plot and the outliers on both sides. On
the other hand has Kygo a normally distributed plot and a lot of
outliers, all on the same side.
The last plot shows the
valence of the songs. This feature is a personal favorite, because of
Spotify’s system for distinguishing what is happy and what is sad music.
In this case it is interesting to see the wide range of Avicii’s songs
and the overall lower score of Kygo compared to Avicii.
Column
Table Comparison
Avicii
Number of songs "49"
Average duration (in seconds) "228.118081632653"
Oldest song "Levels (2011)"
Newest song "Forever yours (Avicii Tribute) (2020)"
Most streamed song "Wake Me Up"
Kygo
Number of songs "77"
Average duration (in seconds) "218.490766233766"
Oldest song "Epsilon (2013)"
Newest song "Whatever (2024)"
Most streamed song "It Ain't Me" Energy
Danceability
Acousticness
Tempo
Valence
Over the years
Column
About the plots
To see the change of the two artists over time, it is necessary to
include the years. One way to do this is to look at each album to see
how this music differs from an album released later. This is done on the
right for Avicii (left) and Kygo (right). Note that a flaw in this
method is the exclusion of songs that are not covered by an album. In
these plots, we can see the valence from 0 to 1 on the x-axis, the
energy from 0 to 1 on the y-axis, the size of the bubbles indicate the
popularity of the track and the color of the bubbles indicate the
danceability (the brighter, the higher, the darker, the lower).
As for Avicii, you can see that almost all the songs are in the
upper left box (low valence, high energy). His most streamed song does
not fall into this. If we then look at his second album, we can see that
just over half are in the upper right box (high valence, high energy).
This looks like an upward trend in valence, but the third album tends to
return to the statistics of the first album. Namely, here the vast
majority are in the upper left box (low valence, high energy). It is
important to keep in mind that Avicii passed away in 2018, while the
album came out in 2019. Therefore, it is possible that his management
have continued with older works from years ago, which could explain the
change back in his music style.
On the right we have the music
of Kygo. In his first album, we again see that the busiest section is
the upper left (low valence, high energy) section. The outliers are now
in the lower left (low valence, low energy). One of them is the song
“Intro,” which is not very representative of the album. In the next
album, the average shifts a bit to the right with some songs in the
upper right box (high valence, high energy). Then, in the third album,
we can see a clear group of bubbles in the upper left box again (low
valence, high energy), with some bubbles scattered around it. In the
last album, we can see a somewhat similar shape, but with a lower
average enegry level. Note that the bubbles in the last plot are pretty
big on average.
We can conclude that for both Avicii as Kygo,
there is not a clear trend that follows through the years, based on
these plots. There are for both of them changes in different albums, but
we can not really make a conclusion of a clear change. These plots are,
however, really interesting to see in my opinion.
Column
True (2013)
Stories (2015)
TIM (2019)
Column
Cloud Nine (2016)
Kids in Love (2017)
Golden Hour (2020)
Thrill Of The Chase (2022)
Self-similarity matrices
Column
About the plots
On the right, we have eight plots. Each song has one chroma-based and
one timbre-based self-similarity matrix. We can see these SSMs of ‘Hey
Brother’ and ‘SOS’ from Avicii and ‘Firestone’ and ‘Higher Love’ from
Kygo (songs can be heard with the displays below). These songs were
chosen because their popularity is not low, which makes them
sufficiently representative. In addition, the release dates are
relatively far apart to allow any differences to be seen.
If
we compare the two songs of Avicii, we can see the same parallel
diagonal lines in the four plots, but the timbre-based SSM seems more
‘blurry’ and the chroma-based SSM has more tiny yellow blocks in it.
The plots of Kygo does not seem to have those parralel
diagonal lines, but they do have a clear structure, which can be seen in
both his songs. This structure resembles the one in ‘Hey Brother’.
Column
Hey Brother (2013)
SOS (2019)
Column
Firestone (2014)
Higher Love (2019)
Chordograms
Column
About the plots
In this section, we have some chordograms of both artists. Again, the
songs are picked on popularity and release year and can be heard down
below.
We can see that the plots all pretty much differ. While
‘Silhouettes’ has a bright ending, ‘The Nights’ and ‘Raging’ has some
parts where multiple chords are bright, while ‘Love Me Now’ has a lot of
them. All of these plots are made with the same method, so we can
conclude that the brighter plots have more variety in chords. The
recurring strips in the plots could be related to build up to a drop or
the drop itself.
Column
Silhouettes (2012)
The Nights (2014)
Column
Raging (2016)
Love Me Now (2022)
Novelty functions
Column
About the plots
On the right, there are two ‘novelty-grams’. One of Avicii’s ‘You
make me’ and the other is Kygo’s ’Stargazing. I set the time to 110 sec,
because that is the part after the first drop for both songs.Here you
can see that the start of Stargazing is calm compared to You make me.
Also, the line of Stargazing is moving somewhat more, whereas the
average line of You make me would be more straight. To be honest, I
expected more differences, because the songs do differ a lot in tempo
and energy.
Note: I wanted to make a tempogram of the songs,
but I kept getting an error, where I had to restart the kernel and I do
not know why this is the case. I got some tips and help, but it still
would not work :(.
Column
You make me - Avicii
Stargazing - Kygo
Clustering
Column
About the plots
The plots on the right show the clusters of the two playlists of
Avicii’s music and Kygo’s. This shows which songs are very different
from each other, where they converge high up in the tree and which are
similar, where they converge low in the tree.
Interesting
findings in Avicii’s tree is that ‘Superlove’ is very different from all
the other songs. The same goes for ‘Somewhere in Stockholm’. In
addition, you notice (if you know the songs) that songs from the same
album are not necessarily close to each other.
As for Kygo’s
tree, similar points can be seen. ‘Intro’ is very different from all the
other songs. This is to be expected because it is a very different song
with also a different purpose in the album. In addition, again, there is
no relationship between songs in the same album and close to each other.
Note: Especially with Kygo’s plot, there are just too many
songs to keep it clear. If you know the names, you can probably
distinguish all the songs. Next to that, I personally find it a lot more
interesting if all the songs are in the tree, instead of clustering per
album. That is why I chose to keep it this way.
Column
Avicii (Hierarchical cluster)
Kygo (Hierarchical cluster)
Conclusion
Column
After analyzing the self-made corpus of Avicii and Kygo we can
conclude a few things.
First of all, there is a clear
distinction between the music of Avicii and Kygo if you listen to their
songs by yourself. However, if we look at all the plots in this report,
I do not think an average person would classify their songs 100%
correctly. Of course, their music genres are close to each other if we
look at all the genres that are available on Spotify. And if you have a
lot of musical knowledge about chords and stuff, maybe it would be
easier to distinguish their songs based on the plots. In the end, I can
conclude for myself that I expected more difference in statistical
terms.
In this report, I also tried to see if I could see a
change over time in their music. I only addressed this topic in the tab
“Over the years”, because of the number of songs I analysed. If I look
at a change over time, I need a lot of songs and this was not really
doable with more complex plots. In this tab however, we already
concluded that there is not really a change to see. It could be that
there is in fact a change existing, but in other features that I have
not addressed or features outside Spotify’s capabilities. For example, I
think that there is a change in Avicii’s lyrics. This is especially
noticeable in his latest album.
Next to the notes I mentioned
within the different tabs, there are some other small notes that are not
really worth mentioning, but did affect this report. Nevertheless, I
hope you enjoyed reading my report. I really enjoyed making it, so at
least someone is happy with it. Analyzing two of my favorite artists is
not something I do everyday. Thanks for reading this report :)